Security audit
Security checks for vulnerabilities and agentic risk
This skill is a lightweight web-research helper with disclosed search and fetch behavior and no evidence of hidden code, persistence, credential access, or destructive actions.
Install this if you want a quick web-research shortcut. Be aware that using it will send search queries about the requested topic to web search/fetch tools, and it is not designed for exhaustive or rigorous fact-checking workflows.
The manifest description says to use the skill when the user asks to "research" something, which is a common everyday request and may overlap with many unrelated contexts. Aside from excluding "deep research," it does not define narrower trigger boundaries or negative examples to prevent unintended invocation.
The "When to Use" section defines activation as when the user says "research ", but does not specify how literal that phrasing must be or what nearby phrasings should not match. This ambiguity can cause accidental invocation from normal conversational requests involving the word "research."
The workflow begins with "When user says "research "" but gives no extra specificity about context, intent, or matching rules. Repeating the broad trigger without constraints increases the chance of inconsistent or over-broad activation.
No suspicious patterns detected.
